410 BELL SYSTEM TECHNICAL JOURNAL 



Thus the effect at x = o is the same as if an electromotive force 



e had been impressed at x = a.^ 



It would now be possible to assume a definite form of disturbance at 

 the surface of the ocean, and by applying the principles that have been 

 discussed in the preceding pages, to work out for any particular cable 

 the wave shape of the resulting interference at the cable terminals. 

 On account of our lack of knowledge as to what might be considered 

 a typical disturbance at the surface of the ocean, such results would 

 be merely, speculative, and would be of no practical value in predicting 

 the actual terminal interference that might be expected. A much 

 better scheme is to compute for each cable, what may be called the 

 interference susceptibility, this being defined, for a particular fre- 

 quency, as the integral 



\A . 6-^* • e-T^- • dx. (5) 



the integration extending over the entire cable. A is a factor which 

 takes account of the shielding by armor wires, and changes at each 

 point on the cable where the armoring changes, z is the depth of im- 

 mersion at a distance x from the terminal, the relation between z and x 

 being obtainable from the profile curve of the cable route. By com- 

 paring the susceptibility-frequency curves for two cables we can obtain 

 an idea of the relative disturbances to be expected on the cables, with 

 the possible exception of that part arising from sources in close proxim- 

 ity to the cables. For the latter type of interference special consider- 

 ations are necessary. 



In drawing conclusions from a susceptibility-frequency curve it is 

 essential to bear in mind that, although the disturbance at the cable 

 terminal is a composite of sinusoidal voltages and currents of all fre- 

 quencies from zero to infinity, we are principally concerned with the 



« An interesting conclusion to be drawn from equation (4) is that the contributions 

 from various portions of a long section of cable due to a uniform disturbance tend to 

 neutralize each other, on account of the fact that they arrive at x = a in various phases. 

 Since y is equal to a+j0, where a is the attenuation constant and /3 the phase con- 

 stant, both per unit length, the quantity t^y can be represented graphically by a 

 vector of length i-^"- and angle ( —5/3). If a were zero the value of the factor l-e-'ST 

 would be zero for 5/3 =o, 2x, Air, 6«-, etc. That is the disturbance picked up over a 



length of cable 5 = -^,n being any integer, would have no effect at the terminal of the 



cable. On account of the fact that a is not zero, the quantity e-^y is less than unity 

 for all the above values of 5 e.xcept s = o, and complete neutralization of the disturb- 

 ance does not occur. In the case of an inductively loaded cable, however, for a given 

 value of o, /3 is many times greater than the value for the corresponding non-loaded 

 cable. This means that neutralization of interference picked up on the loaded cable 

 is much more complete than in the case of a non-loaded cable. 



